Outwitting a Computer
It’s important to remember that a computer, despite all its power, speed and capability, is really a dumb plastic and/or metal box filled with equally dumb electronic widgets. It can do what it’s been told to do, but it can’t think. The thinking has to be done by the user.

The need for such thinking is often apparent in searching a collection database for a particular descriptive word among hundreds, thousands or tens of thousands of entries.
I recall an instance, not very long ago, when I wanted to find a photograph of a pair of brick gatepost columns in a collection of some 23,000 artifacts, photographs and documents.
Not knowing whether the description of this photograph might have referred to “gatepost,” “gateposts,” “gate post” or “gate posts,” I cleverly (so I thought) ran a search query for “gate” with the expectation that this would cover all the possibilities I had envisioned.

To my surprise, the machine delivered far more “hits” than I had expected – and far more than the number of gates, or pictures of gates, that could possibly be in the collection. The computer hadn’t delivered the information that I wanted, but it had done exactly what I asked of it. The collection, as it turned out, included quite a few objects whose shape had been described as “elongate.” There were several “gate valves” among a small lot of antique plumbing fixtures. There was a bottle of patent medicine whose label claimed its ability to “mitigate” the discomfort of Catarrh; a printed card designed to help a Latin student “conjugate” verbs; a letter asking whether acceptance of a corporate grant would “obligate” an institution to display a corporate logo in its announcement of the grant-supported program; a newspaper clipping concerning a court trial and identifying the opposing attorneys who would “litigate” the case; and several items relating to Harvard College commencement exercises and identifying the times and places where various participating groups were to “congregate.”

“Oh $#%^&” I thought, as I set about to outwit the machine. This time I’ll search for “ gate” – i.e., the string I want, preceded by a blank space. This would weed out the spurious “elongate,” “mitigate,” “conjugate,” “litigate,” “obligate” and “congregate.” The “gate valves” would still be found, but I could easily ignore them. I had won — this time!

That’s just one example that might be cited of the occasional need to work one’s way around the sometimes frustrating literal-mindedness of a computer that insists on doing exactly what you asked it to do, when that may not have been exactly what you wanted it to do. It’s often useful, before undertaking a database query, to think about how the machine (which has not a scintilla of common sense) will process the task you’ve assigned to it. That can help to avoid many “Oh $#%^&” moments, and the time required to conceive and process revised queries.

 

Image courtesy of the Computer History Museum

Image courtesy of the Computer History Museum